Priority Substances List Assessment Report for Releases from Primary and Secondary Copper Smelters and Copper Refineries - Releases from Primary and Secondary Zinc Smelters and Zinc Refineries

2.0 Summary of Information Critical to Assessment of "Toxic" under CEPA 1999 (Continued)

2.1 Identity

2.1.1 Definitions and scope

2.1.1.1 Smelters and refineries

For the purposes of these assessments, a smelter is defined as a facility that uses high-temperature chemical processes to recover base metals (MAC, 1995). A refinery is understood to be a facility in which impurities are separated from metals using thermal or electrolytic processes (MAC, 1995). In these assessments, both electrorefineries and electrowinning facilities are considered to be refineries.

Since several metals are recovered from individual smelters and refineries, for the purposes of these assessments a "copper" smelter or refinery is understood to be a facility in which one of its primary commercial products is more or less pure copper metal. Similarly, a "zinc" smelter or refinery has more or less pure zinc metal as one of its primary products.

Primary smelting and refining produce metal directly from ores and concentrates, while secondary smelting and refining produce metal from scrap and/or process waste (Environment Canada, 1997b). The distinction between primary and secondary smelting is not always clear in practice, however, since some predominantly primary smelters use recycled metals to supplement their primary feed.

In a typical copper smelter, a sulphide concentrate or calcine is heated with fluxing agents to about 1200° C, to effect a phase separation into a molten sulphide matte containing copper and iron and an overlying molten slag containing iron oxide, silica and lime. The matte is then subjected to converting and fire refining to produce an impure copper metal known as anode copper, which contains about 99% copper, and minor and trace elements (Skeaff, 1997). Casting of the impure metal into anodes, the configuration suitable for electrorefining, may take place in either the smelter or refinery. A copper refinery then electrolytically refines the anode copper to produce pure copper.

Zinc may be produced using either a roast-leach-electrowin (RLE) process or a pressure-leach-electrowin (PLE) process. Roasting refers to the heating of concentrate to oxidize and drive off sulphur oxide gases. In roasters, zinc and iron sulphides in the concentrate are converted to oxides, and the resulting solid product (calcine) is sent to leaching. Leaching can be under acidic conditions, neutral conditions or a combination. Zinc and non-ferrous metals are extracted, producing a zinc sulphate leach liquor. In the pressure-leach process, zinc concentrate is reground and agitated in an autoclave with oxygen and sulphuric acid. Iron and zinc sulphides are thereby converted to iron and zinc sulphate and dissolved in the leachate. Leach liquor from either the roast-leach or pressure-leach process is purified and directed to electrowinning. In electrowinning, an electric current is passed through the purified liquor, causing the zinc sulphate to chemically decompose and zinc metal to deposit on the cathode. Cathodes are stripped mechanically, and the zinc is melted and cast.

Since the RLE process is a high-temperature chemical process, facilities using it to recover zinc may be considered to be, at least in part, "smelters" as defined previously. However, zinc production facilities may also be considered "refineries" as defined previously, in that they include an electrowinning step. Because of the ambiguity concerning their classification, these facilities are commonly referred to as "zinc plants". The term "zinc plants" is used in this report to describe facilities involved in recovering zinc using RLE or PLE or a combination of the two.

2.1.1.2 Releases

For the purposes of these assessments, a release is considered to be any current discharge to the ambient environment. Past releases, which were often larger and less well controlled than at present, are not included.

Releases considered in these assessments include all current on-site discharges to air and water from Canadian copper smelters and refineries and zinc plants. Releases to air include emissions from "point" (e.g., tall stack) and "area" sources (e.g., low stacks or fugitive emissions from concentrates stored on-site). Effluents considered include both process and cooling waters that are entering surface waters either directly or indirectly (e.g., after passage through a municipal water treatment plant).

As noted in Section 1.0, releases to water from copper smelters and refineries and zinc plants that will be included in effluents regulated under the revised Metal Mining Effluent Regulations (MMER) of the Fisheries Act are not examined in these assessments. Releases from off-site activities related to copper and zinc smelting and refining (e.g., releases from the shipment of feed materials or wastes and landfilling of wastes) were also not considered in these assessments.

Direct impacts of the storage of smelter or refinery wastes (e.g., slags) on lands within the boundaries of facilities were also not examined, since land owned by the facilities is not considered part of the ambient environment. However, leachate or runoff from such wastes that enters ambient off-site waters and wind-blown fugitive emissions from such wastes that are transported off-site were in principle included.

2.1.2 Facilities included in assessments

All copper smelters and refineries and zinc plants currently operating in Canada were included in these assessments.³ Using the definitions presented in Section 2.1.1, six copper smelters, four copper refineries and four zinc plants were identified (Tables 1 and 2). Those with effluents that will be regulated under the revised MMER of the Fisheries Act are identified in these tables. As noted in Section 1.0, risks associated with direct releases to water from these facilities were not assessed. Screening-level assessments of the risk to the environment of aquatic releases were conducted for the Noranda-Canadian Copper Refinery (CCR), Noranda-Canadian Electrolytic Zinc (CEZinc) and Cominco-Trail Operations (CTO).

The Falconbridge-Kidd Creek and HBM&S copper smelters are primary smelters. No currently active stand-alone secondary copper smelters were identified. However, a relatively small portion of the feed entering the Noranda-Horne facility, and to a lesser extent the Noranda-Gaspé facility, is recycled copper-bearing material (Hatch Associates, 1997). These smelters could be considered to be predominantly primary copper smelters that engage in some secondary smelting.

Two primary nickel/copper smelters were also assessed. They are the Falconbridge-Sudbury and Inco-Copper Cliff plants. The Inco plant produces impure copper as well as nickel products. The Falconbridge operation produces only a nickel/copper matte, which is shipped to Norway for further processing (Environment Canada, 1997b). The Falconbridge smelter has been included in these assessments because the operations carried out at Sudbury are the first step of a smelting process that ultimately leads to the production of copper metal.

Table 1 Copper production facilities whose releases were assessed

Enlarge table

Of the four copper refineries identified, three are electrorefineries (Noranda-CCR, Falconbridge-Kidd Creek and Inco-Copper Cliff) and one is an electrowinning plant (Inco-Copper Cliff). Throughout the balance of this report, the Inco-Copper Cliff copper refineries will be referred to as a single operation.

Of the four zinc plants identified, one uses a RLE process (Noranda-CEZinc), one uses a PLE process (HBM&S), and two use both processes (Cominco-Trail and Falconbridge-Kidd Creek). All process only concentrates from zinc ores and hence are "primary" plants. No secondary zinc production plants were identified in Canada.

2.1.3 Release constituents examined

Constituents of releases to water considered in these assessments include all metal contaminants reported to be present, as well as selenium (Se), fluoride, ammonia and pH (hydrogen ion activity).

The components of releases to air that were examined most closely are SO₂, particulate matter (PM) and seven metals (copper-Cu, zinc-Zn, nickel-Ni, lead-Pb, cadmium-Cd, chromium-Cr and arsenic-As⁴). These include the vast majority (on a mass basis) of substances released to air from Canadian copper smelters and refineries and zinc plants (e.g., NPRI, 1995, 1996; RDIS, 1995). Past emissions of both SO₂ and several of these metals from Canadian copper smelters and refineries and zinc plants have been reported to cause environmental harm (e.g., Sanderson, 1998). Hexavalent Cr compounds, inorganic As compounds, inorganic Cd compounds, and oxidic, sulphidic and soluble inorganic Ni compounds were assessed under PSL1 and were found to be CEPA toxic. It should be pointed out, however, that these assessments were not specific to copper smelters and refineries and zinc plants - they considered all sources of entry of the compounds into the environment and therefore do not on their own satisfy the mandate of the current assessments. Other components of releases to air that were examined in the environmental assessment are carbon dioxide (C_O2), nitrous oxide (N₂O) and volatile organic compounds (VOCs).

Table 2 Zinc production facilities whose releases were assessed

Facility	Type	Location	Zinc production (tonnes/year)	Year	Effluent subject to MMER	Source of production data
Cominco	RLE/PLE	Trail, B.C.	264 000	1995	no	A
Noranda-CEZinc	RLE	Valleyfield, Que.	223 000	1995	no	A
Falconbridge-Kidd Creek	RLE/PLE	Kidd Creek (Timmins), Ont.	131 000	1996	yes	A
HBM&S	PLE	Flin Flon, Man.	093 000	1995	yes	B

1. Environment Canada, 1997b.
2. Hatch Associates, 1997.

Among the substances reported in releases to the atmosphere from Canadian copper smelters and refineries and zinc plants that were not examined in these assessments are mercury (Hg) and, in the case of at least one copper smelter, dioxins and furans (Environment Canada, 1997b). While it is recognized that releases of such substances have the potential to harm the environment and human health, their fate in the environment (including accumulation pathways in organisms) is complex and uncertain. The decision not to consider such substances in these assessments was in part a practical one, taking into consideration the anticipated uncertainties associated with estimating their fate (including long-range transport, bioaccumulation and biomagnification) in the environment.⁵ As noted in Section 3.1.1.1.3, this decision contributes to the uncertainty of the overall risk characterization. Polychlorinated dibenzodioxins and polychlorinated dibenzofurans were on the first PSL and were found to be CEPA toxic. Mercury is also on the list of CEPA toxic substances (Schedule I). As pointed out above, however, these conclusions were not based specifically on copper smelters and refineries and zinc plants as the sources of entry to the environment.

2.2 Entry characterization

Voluntary questionnaires were sent to industry in 1998 to verify and update existing information on the chemical constituents of releases, the amounts (expressed as rates) of substances released, the conditions of release (e.g., stack heights and temperatures), the physical and chemical forms of substances released, and concentrations in waste streams (e.g., effluents) and in environmental media near Canadian copper smelters and refineries and zinc plants. Other information collected included the configuration of effluent waste streams and source apportionment of emissions for facilities having multiple operations.

The most recent empirical data available for a complete calendar year were generally used for estimating exposure to metals and ambient SO₂. Data for 1995 were used for atmospheric metal dispersion modelling and for SO₂ source-receptor modelling. All empirical data used in these assessments are for 1995 or a more recent year.

2.2.1 Releases to air

2.2.1.1 Sulphur dioxide

Information on emissions of gaseous SO₂ from copper and zinc facilities is summarized below. Data described are for 1995, since these were the most recent available at the time of information collection. Further detail is provided in SENES Consultants (1999a).

Approximately 99% of the SO₂ emissions from copper and zinc facilities in 1995 were derived from copper smelters, as compared to 1% from copper refineries and zinc plants (see Table 3). Approximately 85% of the SO₂ emissions from these facilities were generated by three copper smelters: Inco's nickel/copper smelter complex in Sudbury, Ontario; Noranda's Horne copper smelter in Rouyn-Noranda, Quebec; and the HBM&S copper smelter in Flin Flon, Manitoba. By 1995, total SO₂ emissions from the copper smelters listed in Table 3 had been reduced by over 61% from 1980 emission levels, and by a total of 63% from copper smelters and refineries and zinc plants as an industry group. Emissions from zinc plants had been reduced by 94%, mainly due to the elimination of SO₂ emissions from the HBM&S zinc plant after 1993. Trend analyses for SO₂ emissions from copper refineries over this period were unavailable because emissions from the Noranda-CCR refinery were incomplete, while emissions from the Inco refinery at Sudbury and the Falconbridge refinery at Kidd Creek were included in the total SO₂ emission inventories from their associated smelters over this period. Although some data on refinery emissions were obtained for the Inco and Falconbridge-Kidd Creek refineries (personal communication with facility operators), the inconsistent consideration of anode casting as either a smelter process or a refinery process complicates their interpretation.

Table 3 Releases of SO₂ to the atmosphere in 1995¹

Facility	SO2 releases in 1995 (tonnes/year)
Copper smelters
Noranda-Horne	169 532
Noranda-Gaspé	043 200
HBM&S-Flin Flon	162 107
Falconbridge-Sudbury	045 000
Inco-Copper Cliff	236 000
Falconbridge-Kidd Creek	005 230
Copper refineries
Noranda-CCR	No data
Falconbridge-Kidd Creek	1 300
Inco-Copper Cliff	0<10
Zinc plants
Noranda-CEZinc	3 300
HBM&S-Flin Flon	0000
Falconbridge-Kidd Creek	0960
Cominco-Trail	1 752

1 Data presented are based on unpublished emissions data from the Residual Discharge Information System (RDIS, 1995), the 1995 annual report for the Eastern Canada Acid Rain Program (Environment Canada, 1995) and additional information provided by facility operators.

In 1995, emissions from copper and zinc facilities accounted for approximately 37% of SO₂ emissions from sources in eastern Canada. However, total SO₂ emissions in eastern North America are dominated by emissions from the United States, with SO₂ emissions in the eastern United States accounting for about 86% of total emissions in eastern North America (Environment Canada, 1997c). Consequently, as a source group, copper and zinc facility emissions represent a much smaller component of total SO₂ emissions in eastern North America. For example, in 1995, SO₂ emissions from Canadian copper smelters and refineries and zinc plants represented only about 5% of the total SO₂ emissions in this region.

2.2.1.2 Metals

The following discussion focuses on metal emissions in 1995, as the dispersion modelling used data from that year. The year 1995 was selected for use as it was the most recent year for which detailed data were available at the time the modelling was begun. No data were identified on the chemical or physical forms of the emitted metals. Most emissions of the metals considered below may, however, be assumed to be in particulate form.

The total annual releases of Cu, Zn, Ni, Pb, Cd and As from the copper and zinc facilities in 1995 are summarized in Table 4. Data in this table are based largely on 1995 data from the NPRI (1995), supplemented with information obtained directly from the facility operators. A more detailed discussion of these releases is provided in SENES Consultants (2000).

The total amount of trace metals released by a facility depends to some extent on the concentration of that element in the raw material fed into the process, as well as on the type of process used, the facility production rate and the efficiency of control equipment at the point of release. If the smelter derives a large proportion of its raw materials from a variety of mining operations, the variability in the concentration of trace elements can result in large fluctuations in trace element release rates, depending on which concentrate is being processed at any given time. Further, for any given trace element, the quantity released in the process exhaust stream will also depend on the temperature of the smelting or the refining process in use. The more volatile elements, such as As, Cd, Pb and Zn, are more likely to be liberated during the process if the temperature is high than are less volatile elements, such as Cu and Ni. Finally, the quantity of each element released is also dependent on the efficiency of the emission control equipment in use at each facility (i.e., multi-cyclones, electrostatic precipitators or baghouses). For all of these reasons, it is not surprising that the data in Table 4 display a high degree of variability in the emission rates among the facilities. The differences range over two to three orders of magnitude for most of the trace metals within each of the three facility categories (i.e., copper smelters, copper refineries and zinc plants).

Among copper smelters, the Inco-Copper Cliff facility ranked highest in 1995 for emissions of trace Cu, Ni and As releases. The Noranda-Horne smelter was a close second for both Cu and As. Noranda-Horne also had the highest emission rates for both Zn and Pb. The highest emission rate for Cd was reported by the HBM&S copper smelter at Flin Flon, followed closely by both Noranda-Horne and Falconbridge-Sudbury.

As with the copper smelters, the trace metal emission rates from copper refineries also vary significantly among the various facilities. As was the case with SO₂ (Section 2.2.1.1), a lack of consistency for inclusion of anode casting as a smelter or a refinery operation complicates interpretation of these data.

Table 4 Releases of metals to the atmosphere in 1995 ¹

Facility	Metal (tonnes/year)
	Cu	Zn	Ni	Pb	Cd	As	Cr
Copper smelters
Noranda-Horne	123	100	1.4	355	4.7	39	1.6
Noranda-Gaspé	1.4	2.8	0.78	17	0.22	16	ND
HBM&S-Flin Flon	62	58	ND	31	6.0	4.5	ND
Falconbridge-Sudbury	9.0	1.9	10.2	13.8	4.5	1.0	0.47
Inco-Copper Cliff 2	132	9.5	87	39	2.4	48	ND
Falconbridge-Kidd Creek	29	4.8	0.2	9.4	0.44	0.78	ND
Copper refineries
Noranda-CCR	2.6	ND	0.027	1.27	ND	0.086	0.00
Falconbridge-Kidd Creek	neg.	neg.	neg.	neg.	neg.	neg.	ND
Inco-Copper Cliff	28	ND	0	0	ND	1.1	ND
Zinc plants
Noranda-CEZinc	0.7	106	ND	0.9	0.9	0.2	ND
HBM&S-Flin Flon	0	0	ND	0.12	0.004	0	ND
Falconbridge-Kidd Creek	0.16	8.3	0.013	0.06	0.03	4.8	ND
Cominco-Trail 3	ND	18	ND	0.25	0.015	ND	ND

ND - Not determined; neg. - negligible

1. Emission values for Cu, Zn, Ni, Pb, Cd and As were used for dispersion modelling. Chromium was considered only in the assessment of risk to human health. Trace metal emission data were based largely on NPRI data (NPRI, 1995) with additional information provided by facility operators. Values have been rounded for presentation.
2. Values shown are based on the average of results from two samplings of the main Inco stack, one from 1994 and the other from 1996. It should be noted that these values differ significantly from those reported to the NPRI in 1995 for the smelter complex (Inco reports releases from the copper refinery separately). NPRI (1995) values for the smelter complex were: Cu-107.04; Zn-15.65; Ni-417.76; Pb-68.23; Cd-ND; As-7.32 tonnes.
3. Cominco-Trail is a combined facility, including both a zinc plant and a lead plant. NPRI data are reported for the overall facility and do not distinguish between the plants. Therefore, the values shown in the table were based on 1995 source attribution data provided by facility operators in response to a questionnaire from Environment Canada. It is believed that the provided list of emission sources reportedly associated with the zinc plant was incomplete, resulting in underestimation of emissions from this operation. For comparison, more accurate metal emission estimates for the Cominco zinc plant in 1998 (personal communication with facility operators) were: Zn-125; Pb-0.36; Cd-0.124; As-0 tonnes. The facility operators point out that the 1998 estimates are based on releases for the point sources associated with each operation (zinc and lead). This is an oversimplification because of the numerous recycle streams between zinc and lead operations. Therefore, there is considerable uncertainty associated with these values.

Emission rates for trace metals from zinc plants also vary by two to three orders of magnitude. Among zinc plants, the Noranda-CEZinc facility had the highest emission rates for Cu, Zn, Pb and Cd. In fact, the Zn emission rate from this facility exceeded that from all other copper and zinc facilities, including the emissions from the Noranda-Horne copper smelter. However, as discussed in the footnote to Table 4, it is believed that the emissions attributed to the Cominco zinc plant in 1995 may have been significantly underestimated. For comparison, Zn emission values attributed to the Cominco zinc plant for 1998 were estimated at 125 tonnes (personal communication with facility operators), greater than those from any of the other copper or zinc facility operations for 1995. The Falconbridge-Kidd Creek plant had the highest emission rate for As among the zinc plants and was the only zinc plant to report Ni emissions, although the rate is very low. Releases of all metals from the HBM&S zinc plant were reported to be very low, since, in contrast to the others, this plant does not use a high-temperature roasting process.

2.2.1.3 Particulate matter

Table 5 summarizes 1995 emissions of total particulate (TP) matter for the copper and zinc facilities as contained in the Residual Discharge Information System (RDIS, 1995). The table also includes emission data for the size fraction less than or equal to 10 mm (PM₁₀) and the fraction less than or equal to 2.5 mm (PM_2.5).

These data indicate that the Inco copper smelter at Copper Cliff had the highest TP emission rate at 7052 tonnes per year in 1995. Note that this total includes TP emissions from the copper refinery and nickel refinery as well as from the smelter. The Noranda-Horne copper smelter ranks second at 1339 tonnes per year, followed by the Falconbridge copper smelter at Sudbury, with 1181 tonnes per year. The lowest reported TP emission rate for the copper smelters is 430 tonnes per year at the Noranda-Gaspé facility.

TP data for two of the three copper refineries are included in the TP emissions from the copper smelters (Inco at Copper Cliff and Falconbridge-Kidd Creek). Only the Noranda-CCR facility is listed as a separate copper refinery source, at 7.1 tonnes per year.

Among the zinc plants, the highest TP emission rates listed in the RDIS (1995) are for the Noranda-CEZinc and Cominco-Trail facilities, both of which emit about 150 tonnes per year. TP emissions from the Falconbridge- Kidd Creek zinc plant are not listed separately and are included with emissions from the copper smelter and refinery located at this site.

TP emissions from RDIS (1995) may be subdivided into three categories (Table 5):

• fugitive sources - consisting of dust from roads, wind erosion of exposed surfaces, and releases from material handling and storage on-site;
• low-elevation sources - consisting of releases from short stacks (defined as less than 30 m high); and
• high-elevation sources - consisting of all releases from stacks greater than 30 m high.

There are some anomalous features associated with these TP emissions. For example, three copper smelters list identical estimates of fugitive emissions of 500 tonnes per year. These appear to be notional numbers, and are unlikely to have been based on detailed emission calculations. In addition, fugitive emissions are not reported for any of the Noranda facilities, or for Cominco-Trail. Because of these anomalies, and the fact that trace metal to TP mass ratios at several sites appear to be unusually high, TP emission data in Table 5 must be interpreted with particular caution. It should be noted that there are some inconsistencies between facilities in the reporting of both TP and metals emissions. Among the sources irregularly reported are fugitive emissions from the production, storage and handling of concentrates, exhaust from baghouses and wind-blown dust from uncovered tailings. The emissions sources reported by four zinc and copper processing facilities have been evaluated and are detailed in SENES Consultants (1999b).

Although data on the size of the particles released were very limited, it is expected that fugitive releases are relatively coarse (>2 mm). Results of preliminary work on particle size distributions of PM obtained from the stack of the Inco smelter indicate that the particles are extremely fine, with 80% less than 3 mm (Burnett, 1998).

Table 5 Releases of total particulate (TP) matter, particulate matter less than or equal to 10 mm (PM₁₀) and particulate matter less than or equal to 2.5 1,2 mm (PM_2.5) to the atmosphere in 1995 ^1,2

Facility	Total particulate (TP)				PM10 (tonnes/ year)	PM2.53 (tonnes/ year)
	(tonnes/ year)	Percent
		Fugitive	Low stack	High stack 4
Copper smelters
Noranda-Horne	1339	-	2	098	1091	0866
Noranda-Gaspé	0430	-	-	100	0301	0117
HBM&S-Flin Flon	0717	48	-	052	0427	0180
Falconbridge-Sudbury	1181	42	6	052	0857	0665
Inco-Copper Cliff 5	7052	07	2	091	6654	5531
Falconbridge-Kidd Creek	0504	99	1	-	0187	0097
Copper refineries
Noranda-CCR	07.1	-	100	-	05.2	04.0
Falconbridge-Kidd Creek	included with copper smelter emissions
Inco-Copper Cliff	included with copper smelter emissions
Zinc plants
Noranda-CEZinc	153	-	100	-	0119	0107
HBM&S-Flin Flon	078	100	-	-	0067	0023
Falconbridge-Kidd Creek	included with copper smelter emissions
Cominco-Trail	156	-	100	-	0134	0115

1. Data obtained from RDIS (1995). Values have been rounded for presentation. As the reliability of some of these data has not been established, there is considerable uncertainty associated with some values.
2. Some emission values include sources that are associated with and reported by the facilities, but which are not the subject of these assessments. Owing to questions on the reliability of the data, more rigorous evaluation of source attribution was not warranted and was in some cases precluded.
3. It should also be noted that PM (mostly in the form of PM_2.5) can form from condensation of smelter gases after release to the atmosphere. Therefore, attribution of ambient PM_2.5 based on emissions could underestimate the proportion due to smelting processes.
4. "High stacks" are defined in this assessment as being more than 30 m in height.
5. TP values for Inco-Copper Cliff include emissions from the smelter complex, the copper refinery and the nickel refinery.

The values for TP emissions discussed above do not take into account the secondary formation of PM. Secondary processes involve the formation of PM (usually PM_2.5) in the atmosphere as a result of physical and chemical transformation of gases. Sulphur dioxide, nitrogen oxides and VOCs are among the major contributors to the formation of PM_2.5 (EC/HC, 2000a).

2.2.1.4 Carbon dioxide, nitrous oxide, methane and volatile organic compounds

Emissions of gaseous C_O2, N₂O, methane (CH₄)⁶ and VOCs from copper and zinc processing facilities are summarized in Table 6. These compounds are of significance due to their influence on abiotic atmospheric effects such as climate change and the formation of ground-level ozone.

As mentioned in the previous section, both VOCs and nitrogen oxides are significant precursors in the secondary formation of PM_2.5. Total emissions of the oxides of nitrogen from the facilities being considered in these assessments were about 1800 tonnes in 1995 (RDIS, 1995).

2.2.2 Releases to water

Information on releases to water from CCR, CEZinc and CTO is summarized below. Further details are provided in Beak International (1999).

Annual average loading rates from the three facilities into their receiving environments are shown in Table 7. Factors applied to annual averages to estimate maximum short-term (monthly and four-day mean) loading rates are summarized in Table 8. These factors are based on empirical loading information. Concentrations of release components in undiluted effluents are shown in Table 9.

2.2.2.1 Canadian Copper Refinery

The waste metal loadings from CCR (Table 7) are released to the Montreal Urban Community (MUC) wastewater treatment plant (WWTP), which discharges in turn to the mid-channel St. Lawrence River east of l'Île aux Vaches. On a volume basis, approximately 7% of the wastewater leaving CCR is treated process water, and the remainder is untreated cooling water drawn from the St. Lawrence River. On a mass basis, Cu and Se are the two most significant loadings. In 1995, these metal loadings comprised 0.82 and 3.58 tonnes, respectively. Loadings of most metals increased significantly in 1996.

The MUC-WWTP removes much of the CCR loading prior to entry into the St. Lawrence River. Typical removal rates at the MUC-WWTP and the total annual loadings of pertinent metals to the St. Lawrence River in MUC-WWTP treated effluent were taken from Deschamps et al. (1998). There is considerable uncertainty in both removal rates and loadings for certain metals, such as As and Se, that are measured at the MUC-WWTP at concentrations close to the analytical detection limit.

The subsequent assessment of biological exposures to metal in the St. Lawrence River and associated effects on aquatic biota must be based upon the loadings from the MUC-WWTP (not CCR), since these are the loadings actually received by the St. Lawrence River. However, it is important for the purposes of this assessment to identify the proportional contribution that CCR makes to the release of metals in MUC-WWTP effluent. This proportion is calculated for each metal as follows:

P_CCR = L_CCR/[(C_MUC/[1- R_MUC])*Q_MUC]

where:

• P_CCR = proportional contribution of CCR
  (fraction),
• L_CCR = metal loading from CCR to MUC-WWTP (mg/s),
• R_MUC = metal removal rate at MUC-WWTP
  (fraction),
• C_MUC = metal concentration in MUC-WWTP effluent (mg/L), and
• Q_MUC = volume of MUC-WWTP effluent (L/s).

Table 6 Releases of carbon dioxide (C_O2), nitrous oxide (N₂O), methane (CH₄) and other volatile organic compounds (VOCs) to the atmosphere in 1995 ¹

Facility	Releases in 1995 (tonnes/year)
	CO2	N2O (as CO2 eq.)2	CH4 (as CO2 eq.)2	VOCs
Copper smelters
Noranda-Horne	105 210	490	36	2.08
Noranda-Gaspé	120 090	1 630	95	1.37
HBM&S-Flin Flon	NR	NR	NR	NR
Falconbridge-Sudbury	NR	NR	NR	1.81
Inco-Copper Cliff	NR	NR	NR	3.03
Falconbridge-Kidd Creek	NR	NR	NR	4.62
Copper refineries
Noranda CCR	80 917	264	43	2.05
Falconbridge-Kidd Creek	NR	NR	NR	included in copper smelter value
Inco-Copper Cliff	NR	NR	NR	included in copper smelter value
Zinc plants
Noranda-CEZinc 42 288		437	21	1.16
HBM&S-Flin Flon	NR	NR	NR	NR
Falconbridge-Kidd Creek	NR	NR	NR	included in copper smelter value
Cominco-Trail	NR	NR	NR	NR

NR - Releases not reported

1. Data obtained from the Residual Discharge Information System (RDIS, 1995).
2. To facilitate their interpretation in terms of potential influence on climate change, values for N₂O and CH₄ have been converted to equivalents of C_O2 using global warming potential multipliers of 310 and 21, respectively (Jaques et al., 1997).

The proportional contribution that CCR makes to the metal loading from the MUC-WWTP to the St. Lawrence River (Table 7) ranges from approximately 0.1% for metals such as Cd and Cr to approximately 1% for Cu and Ni, 10% for As and approaching 100% for Se.

Most of the metals released at the MUC-WWTP outfall are presumed to be in dissolved or adsorbed form. The total metal release was assumed to be available to partition with suspended solids in receiving water, and to contribute accordingly to exposures of aquatic biota.

Temporal variability in the MUC-WWTP loadings is uncertain. Neither monthly mean nor daily loading data were available.

Table 7 Annual loading rates of effluent components (tonnes/year)

Enlarge image

Table 8 Factors applied to annual effluent loadings to estimate maximum short-term loading rates (maximum monthly and four-day mean loadings)

Parameter	Averaging period	Ratio of maximum short-term mean to annual mean loading rates
		Noranda-CEZinc 1	Cominco-Trail 2
			C-II outfall	C-III outfall
Cu	1 month	3.9	2.03	1.79
	4 days	-	6.40	3.42
Zn	1 month	1.7	2.13	2.00
	4 days	4.76	3.92	5.54
Pb	1 month	-	2.57	1.39
	4 days	-	5.99	3.15
Cd	1 month	2.4	1.87	1.43
	4 days	-	6.82	3.98
As	1 month	-	1.38	1.73
	4 days	-	2.85	5.41
Hg	1 month	6.2	2.95	2.08
	4 days	-	5.94	5.37
Se	1 month	2.8	-	-
	4 days	-	-	-
Tl	1 month	-	-	3.09
	4 days	-	-	319.30 3
Ammonia	1 month	-	-	1.28
	4 days	-	-	2.26
Fluoride	1 month	-	-	1.13
	4 days	-	-	1.67

1. Noranda-CEZinc data (1995).
2. Cominco-Trail data (1998).
3. Driven by a Tl upset event in April 1998.

2.2.2.2 Canadian Electrolytic Zinc

The waste metal and ammonia loadings from CEZinc (Table 7) are released to the Beauharnois Canal in the St. Lawrence River. On a volume basis, approximately 4% of the wastewater leaving CEZinc is treated process water, and the remainder is untreated cooling water drawn from the Beauharnois Canal. On a mass basis, for the combined effluent, ammonia, Zn and Se are the three most significant loadings. In 1995, these loadings comprised 24.0, 3.24 and 2.5 tonnes, respectively. Process changes in 1999 have resulted in significant reductions in Se loadings from 1995 levels.

Table 9 Concentrations (mg/L) of metals and other constituents in undiluted effluents

Release component	Noranda-CCR 19952	Noranda-CEZinc 1995 combined 3	Cominco-Trail - 1998 4
			C-II outfall	C-III outfall
Cu
Annual mean	31	3.8	8.6	11.0
Maximum 1-month mean	-	14.9	19.0	20.0
Maximum 4-day mean	-	-	43	31.1
Mean % dissolved/adsorbed 1	-	-	65	62
Zn
Annual mean	-	51.4	1194	390.3
Maximum 1-month mean	-	91.6	2600	767
Maximum 4-day mean	-	258	3615	1781
Mean % dissolved/adsorbed 1	-	91	38	59
Ni
Annual mean	9	-	-	-
Mean % dissolved/adsorbed 1	-	-	-	-
Pb
Annual mean	4	2.2	84.6	122.7
Maximum 1-month mean	-	-	225	166.9
Maximum 4-day mean	-	-	392	319
Mean % dissolved/adsorbed 1	-	-	23	34
Cd
Annual mean	0.5	0.21	7.9	7.4
Maximum 1-month mean	-	0.54	15.1	10.4
Maximum 4-day mean	-	-	41.7	24.2
Mean % dissolved/adsorbed 1	-	-	80	81
As
Annual mean	1	-	6.0	20.4
Maximum 1-month mean	-	-	8.8	36.0
Maximum 4-day mean	-	-	13.2	91.0
Mean % dissolved/adsorbed 1	-	-	89	79
Cr
Annual mean	7	-	-	-
Mean % dissolved/adsorbed1	-	-	-	-
Hg
Annual mean	-	0.03	0.4	1.0
Maximum 1-month mean	-	0.24	1.2	2.1
Maximum 4-day mean	-	-	1.7	4.6
Mean % dissolved/adsorbed	-	-	-	-
Se
Annual mean	3	39.7	-	-
Maximum 1-month mean	-	106	-	-
Mean % dissolved/adsorbed 1	-	-	-	-
Ag
Annual mean	2.7	-	-	-
Mean % dissolved/adsorbed 1	-	-	-	-
Tl
Annual mean	-	-	0.0	94.8
Maximum 1-month mean	-	-	-	306
Maximum 4-day mean	-	-	-	1505
Mean % dissolved/adsorbed	-	-	-	94
Ammonia
Annual mean	-	381	0.0	1175
Maximum 1-month mean	-	-	-	1232
Maximum 4-day mean	-	-	-	1843
Mean % dissolved/adsorbed	-	-	-	100
Fluoride
Annual mean	-	-	0.0	2457
Maximum 1-month mean	-	-	-	2703
Maximum 4-day mean	-	-	-	3376
Mean % dissolved/adsorbed	-	-	-	100

1. Percentage of total concentration that is dissolved, plus the adsorbed portion of that which is particulate, according to K_d and suspended solids in effluent.
2. Annual means are based on weekly composite samples (Deschamps et al., 1998). Maximum monthly and 4-day average concentrations have been estimated from maximum monthly or 4-day mean loading and flow for the corresponding period.
3. Concentrations calculated as annual loading (NPRI) ÷ annual discharge, based on weekly composite samples analysed in two effluent streams. Maximum monthly and 4-day average concentrations have been estimated from maximum monthly or 4-day mean loading and flow for the corresponding period.
4. Concentrations calculated as mean daily loading (Cominco data) ÷ mean daily discharge.

The treated process water (UNA) effluent and the cooling water (Principal) effluent have historically been discharged at separate points, about 1 km apart, on the Beauharnois Canal. However, they are now being (or will soon be) released together at the Principal effluent location. For the purpose of the assessment of biological exposures to effluent constituents in the Canal and associated effects on aquatic biota, the two effluents are considered here together as a combined effluent at a single point of release. As compared to a separate UNA discharge, this makes for a larger point source loading, lower end-of-pipe concentrations and less rapid near-field dilution.

Virtually all of the ammonia released at CEZinc will be dissolved, and most of the metal released will be in dissolved or adsorbed form (labile). However, some 5-10% of the zinc may be in a fine particulate metal oxide or hydroxide form. This conclusion is based on CEZinc observations that approximately 60% of Zn in the UNA effluent is not dissolved. With an average 13 mg/L of suspended solids, we might expect 15% of Zn to be adsorbed based on distribution coefficients (Beak International, 1999), but the remaining 45% of Zn in UNA effluent must have a more integral association with PM. Metal oxide particles, formed in the roasting process, are unlikely to dissolve later if released. Metal hydroxide particles, formed in the water treatment process, may dissolve later, although slowly. Here it is assumed that 20%x45% = 9% of the total Zn loading is in such relatively inert forms. Only the portion of loading estimated to be in dissolved or adsorbed form (91%, Table 9) was considered to be available to partition with suspended solids in receiving water, and hence to contribute to exposures of aquatic biota.

Temporal variability in metal loading from CEZinc (Table 8) is based on 1995 monthly composite data for most metals and on daily data, which were available only for Zn. Maximum monthly average loadings range from 1.7 times the annual average to 6.2 times the annual average, depending on the metal. Factors for 4-day average loadings would be higher.

2.2.2.3 Cominco Trail Operations

CTO includes zinc and lead refinery operations, as well as a fertilizer plant. There are three main combined effluent outfalls that contribute chemical loadings to the Columbia River, as well as some residual drainage from a historical landfill area via Stoney Creek. Most of the landfill drainage toward Stoney Creek is now collected and treated.

The Combined IV outfall (C-IV) and Stoney Creek are furthest upstream (Figure 1). The C-IV outfall, associated with the fertilizer plant, is the dominant source of ammonia, but a minor source of metal loadings. Stoney Creek is a significant source of metal loadings.

The Combined III outfall (C-III), approximately 1.3 km downstream from C-IV, is primarily associated with zinc operations. It is a source of ammonia and a significant source of metals and fluoride (Table 7). The C-II outfall, a further 0.8 km downstream, includes contributions from zinc sulphide leaching, as well as lead and other operations. It, too, is a significant source of metals.

A blast pond discharge (C-I) once existed further downstream, and was a very minor source of metal loadings, primarily associated with lead operations. This discharge no longer exists.

Recent (1998) loadings from CTO and specifically from the C-II and C-III outfalls are summarized in Table 7, along with 1995 and 1996 loadings for all CTO as reported in NPRI (1995, 1996). The average 1998 loadings from C-II and C-III were utilized in this assessment to represent the zinc smelting/refining operations at CTO.

The proportion of loading attributed to zinc operations (Table 7) was estimated for each metal in each outfall, based on an evaluation of toxic unit contributions from different sewers to each outfall (Duncan and Antcliffe, 1996). For C-II, only sewer #6 is associated with zinc operations (it drains the zinc sulphide leaching plant and cadmium plant). For C-III, all contributing sewers are associated with zinc operations, except for the contributions from the effluent treatment plant, which were apportioned to lead and zinc operations based on inflow from these areas (60% from zinc operations). Table 7 shows that most of the C-III loadings (77-99% for metals) are related to zinc operations, while for C-II the proportion ranges from 25-91%, depending on the metal.

Figure 1 Map of Trail, B.C., showing the outfalls and sampling locations considered in the assessment of aquatic releases from the Cominco facility

Virtually all of the ammonia and fluoride released from CTO will be in dissolved form. However, a significant portion of the metal loading, particularly for Zn and Pb, is evidently not in dissolved form, based on analysis of filtered and unfiltered effluent samples. With an average of 2-3 mg/L suspended solids in these samples, we might expect up to 5% of some metals to be adsorbed based on distribution coefficients (Beak International, 1999), but this cannot account for all of the undissolved fraction. Thus, the undissolved fraction may be substantially composed of metal oxides, hydroxides or other relatively inert forms. Only the portion of loading estimated to be in dissolved or adsorbed form (as shown in Table 9) was considered to be available to partition with suspended solids in receiving water, and hence to contribute to exposures of aquatic biota.

Temporal variability in chemical loadings from CTO (Table 8) is based on 1998 daily data for all chemicals. Maximum monthly loadings range from 1.2 times the annual average to 3 times the annual average, depending on the metal and outfall. Four-day average loadings may be as high as 2.2-6.8 times the annual average, depending on the metal and outfall.

2.3 Exposure characterization

2.3.1 Releases to air

For releases to air, exposure is quantified both as concentrations in ambient air and as rates of deposition from air. Results for both empirical monitoring and model calculations are presented when available.

2.3.1.1 Sulphur dioxide

2.3.1.1.1 Fate of sulphur dioxide in air

The fate of SO₂ released to air, discussed briefly here, is considered in more detail in SENES Consultants (1999a).

The conversion of SO₂ to sulphate (SO₄^2-) and its subsequent deposition are governed by a complex series of interactions that include transport and diffusion (i.e., dispersion), gas-phase and aqueous-phase chemistry, meteorology, cloud physics, and dry and wet scavenging processes. Field studies of oxidation rates in clouds suggest that aqueous-phase oxidation mechanisms are considerably faster at converting SO₂ to SO₄^2- than gas-phase reactions. Aqueous- phase oxidation of SO₂ in the atmosphere can occur in cloudwater, fogwater and rainwater, in deliquescent aerosol droplets at high relative humidity and in the liquid surface film condensed on aerosol particles (Radojevic, 1992).

Sulphate formed from oxidation of SO₂ often takes the form of fine PM (PM_2.5). For example, for samples collected at 14 urban sites in the National Air Pollution Surveillance (NAPS) network operating from 1986 to 1994, an average of 17% and up to 95% of PM_2.5 collected at each site was composed of SO₄^2- (Brook et al., 1997, as summarized in EC/HC, 2000a).

According to Hidy (1994), SO₂ concentration patterns tend to be consistent with the observed distribution of SO₄^2- concentrations (as a secondary pollutant) in air and rainwater, except that SO₄^2- is more widely and uniformly distributed than SO₂ due to the time required for 2-the transformation of SO₂ to airborne SO4 (about 1 day) and the efficient scavenging of SO₄^2- by precipitation.

Regional airborne SO₄^2- concentrations exhibit episodic behaviour which is linked to stagnant high-pressure systems (Hidy, 1994). In general, high SO₄^2- concentrations in eastern North America are associated with high temperatures, high absolute humidity, moderately low atmospheric pressures and wind speeds, and high ozone concentrations indicative of high oxidation potential. Thus, the year-to-year spatial variability of high SO₂ and SO₄^2- episodes is closely related to the path and frequency of migratory high-pressure systems. Regional high SO₄^2- episodes can occur in all seasons and are positively correlated with temperature, except in winter, when low temperatures associated with near-surface inversions and poor ventilation can also result in high SO₄^2- pollution episodes. In eastern North America, regional SO₄^2- episodes typically last up to 5 days, but extended events may last up to 11 days. Monitoring studies in eastern Canada suggest that as much as 60-70% of the total annual wet SO₄^2- deposition may be deposited in a single deposition episode (Environment Canada, 1997c).

In contrast, the critical factors for dry deposition are the concentration of SO₂ in the air very near the surface and the ability of the surface to "capture" pollutants that come into contact with it (Hicks, 1992). Dry deposition rates are characterized by a strong diurnal cycle and are intrinsically linked to ambient air concentrations. Very little deposition occurs at night, while high rates of daytime deposition are dominated by the frequency and intensity of high air pollution episodes. Recent research on dry deposition suggests that, on average, dry deposition accounts for about 25% of total (dry plus wet) sulphur deposition, although considerable uncertainty remains in the estimates of dry deposition values (Environment Canada, 1997c). Furthermore, the relative magnitude of dry deposition contributions to total sulphur deposition varies seasonally, and the relative importance of wet versus dry deposition varies with location. Dry deposition is relatively more important than wet deposition in areas close to major source regions, while wet deposition dominates at greater distances, indicating the importance of long-range transport processes. Of the total sulphur emitted in eastern North America, it is estimated that about one-third of the annual average amount of anthropogenic sulphur emissions is wet-deposited in eastern North America, while two-thirds is either dry-deposited or transported out of this region.

2.3.1.1.2 Concentrations of sulphur dioxide in air

This section includes a brief description of the analytical and statistical methods used to estimate SO₂ concentrations in air at monitoring stations located close to copper smelters and refineries and zinc plants. The approach used for source attribution is also described. Resulting data on SO₂ concentrations in air at these sites are summarized in Tables 10, 11 and 12.

Monitoring methods: Monitoring of ambient SO₂ levels is conducted using instrumental methods such as fluorescence detectors. These instruments work on a continuous basis or with a high sampling frequency, such as one reading per minute. The detector signals are averaged over some period of time (typically 5-15 minutes), and these values are recorded. The recorded levels are averaged over longer time periods (usually 1 hour and 24 hours) to determine whether permit levels and provincial regulations are being met.

Data sources and analysis: Data were generally obtained from the facility operators as 1-hour averages. These were then used in the calculation of 24-hour and monthly average ambient SO₂ concentrations. Ambient concentrations averaged over the growing season were generally calculated from monthly averages. The growing season was defined to include the months April through October. One-hour and growing season averages were used in the assessment of risk to the environment. Twenty-four-hour averages were used in the assessment of risk to human health.

Environmental assessment: Handling of values below the detection limit ("non-detects") is a significant issue when dealing with ambient SO₂ data, for two reasons. First, up to 98% of the data may be non-detects. These values therefore exert a significant influence on temporal averages. Second, most monitoring is performed to ensure that relatively high, short-term thresholds are not exceeded. Therefore, monitoring is conducted or data are recorded to levels that are less sensitive than may be of significance when considering chronic exposure.

The normal method of correcting for non-detects is to set all values for non-detects to one-half the detection limit. This makes the assumption that non-detect values are equally distributed between zero and the detection limit. Detection limits for data used in this assessment ranged from 0.5 to 50 mg/m³. Setting non-detects to one-half the detection limit would have raised some estimates of seasonal average concentration by close to 25 mg/m³. This would have caused the chronic effects threshold values for environmental organisms to be exceeded owing to non-detected concentrations. To preclude this, only non-detects having associated detection limits of less than 3 mg/m³ were corrected to one-half the detection limit. One exception to this was a facility having several monitors with detection limits of 13 mg/m³. Non-detect values from these monitors were corrected to one-half the detection limit, as the percentage of non-detects was low, resulting in their having only a minor influence on seasonal averages.

Table 10 Growing season average SO₂ concentrations at monitoring stations located near copper and zinc production facilities

Enlarge image

Table 11 One-hour average SO₂ concentrations during the growing season at monitoring stations located near copper and zinc production facilities

Enlarge image

Table 12 Annual summary of 24-hour ambient air concentrations of SO₂ near copper smelters and refineries and zinc plants in Canada

Enlarge image

Correction for non-detects when estimating chronic (i.e., growing season) exposures using data from monitors having higher detection limits was conducted as follows. This includes all monitors at the Falconbridge-Kidd Creek, Noranda-Horne and Noranda-Gaspé facilities, which had detection limits of 25-50 mg/m³. Seasonal averages for these sites were estimated using a statistical method (El-Shaarawi, 1989; El-Shaarawi and Esterby, 1992). This involved fitting detectable values to a suitable statistical distribution; using this distribution to estimate concentrations for sample values that were below the detection limit; and, finally, calculating the average using all detected and estimated values. The Weibull distribution was found to best describe the data (El-Shaarawi, 1999). It has the form:

F(x) = 1 - Exp{-[(x - x₀)/k]^m}

where x₀ is the regional background ⁷ concentration, which is taken to be 2.6 mg/m³ (Linzon, 1999), and "m" and "k" are fitting parameters for shape and scale, respectively. Further information on the methods used is provided in CED (2000).

Growing season average concentrations are shown in Table 10. The percentages of values that were above the detection limit ("% detects") are also indicated. Values shown in bold are above the chronic Estimated No-Effects Value (ENEV) (10 mg/m³), and those bolded and underlined are above the Critical Toxicity Value (CTV) (21 mg/m³) for sensitive vegetation. These effects thresholds are discussed in Section 2.4.1.1.1. One-hour average concentrations are shown in Table 11. In this table, the number of 1-hour averages measured over the period of the growing season that fall into defined concentration ranges is indicated. The concentration intervals shown correspond to the acute (1-hour) ENEV (450 mg/m³) and CTV (900 mg/m³) for sensitive vegetation (discussed in Section 2.4.1.1.1). A more detailed description of these data and their processing is contained in CED (2000).

Human health assessment: The health assessment was based on data for the entire year, rather than being restricted to the growing season. Summary statistics for the 24-hour averages over the most recent year for which data were provided are summarized in Table 12 for all of the facilities except for Noranda-CCR, for which data were not obtained. For each site, the table includes the arithmetic mean and maximum concentration of SO₂ for the most recent year for which data were provided, as well as the identity, location and type of site (e.g., residential) and the number of samples. The frequency of samples with concentrations in various ranges, corresponding to the 24-hour WHO Air Quality Guideline for Europe for SO₂ of 125 m g/m³ (WHO, 1987, 2000), is also presented for each site.

The arithmetic mean values in Table 12 were calculated by assuming a value of one-half of the detection limit for those samples that did not contain detectable levels of SO₂. This assumption can affect the mean markedly when detection limits are relatively great (as for Falconbridge-Kidd Creek, Noranda-Gaspé and Noranda-Horne) and ambient concentrations relatively low. Those instances when the mean value is affected by more than 30% by this assumption are indicated in the table.

Near those facilities with multiple monitoring sites, the arithmetic mean 24-hour concentration of SO₂ is generally increased in relation to the proximity to the facility. In addition, the mean level of SO₂ at virtually all of the sites is elevated compared to background levels at remote or rural locations, which are reported to be 5 mg/m³ or less (FPACAQ, 1987; Linzon, 1999; WHO, 2000). As noted in the table, these increased levels are also reflected in exceedences of the 24-hour WHO Air Quality Guideline for Europe for SO₂ of 125 m g/m³.

Source attribution: Both Table 10 and Table 11 include a column labelled "Emission-based source attribution." These attributions are based on the percentage contribution of separate operations to total emissions from a facility. In relating these source attributions to monitored concentrations, it is assumed that the amount that each source is contributing to measured ambient SO₂ concentrations is proportional to its emissions. For facilities comprising only a copper smelter or copper refinery or zinc plant, 100% of SO₂ emissions from the facility may be attributed to those sources. However, for combined facilities, such as those at Sudbury, Kidd Creek or Trail, several different operations may be contributing to total SO₂ emissions.

For these combined sources, the percentage contribution of each operation of concern to total SO₂ emissions from the facility has been estimated based on the following:

• The Sudbury region includes the Inco facility as well as the Falconbridge facility. As they share an airshed, results for ambient SO₂ monitoring in this region are influenced by the presence of the two. The Inco facility includes a nickel/copper smelter, a copper refinery and a nickel refinery, while the Falconbridge-Sudbury facility includes only a nickel/copper smelter. Source attribution was determined using a combination of emission data for 1995 contained in MacLatchy (1996) and from personal communication with Inco facility operators. Due to a lack of data on emissions from the Inco nickel refinery, it was assumed that SO₂ emissions were - like those from the Inco copper refinery -negligible. No major process changes at either of these facilities are believed to have occurred since 1995.
• The Falconbridge-Kidd Creek facility includes a copper smelter, copper refinery, zinc plant and concentrator. Source attribution was determined from 1995 emission data provided by Falconbridge. No major process changes are believed to have occurred at this facility since 1995.
• The Cominco-Trail facility includes a lead smelter and a zinc plant. The fraction of emissions attributable to the zinc plant was determined based on 1998 emission data provided by Cominco (personal communication with facility operators). These data reflect the significant process changes that took place at the facility in 1996.
• The HBM&S facility in Flin Flon includes a zinc plant and a copper smelter. Due to the process used, the zinc plant does not emit SO₂. Therefore, all of the SO₂ detected is attributed to the copper smelter, and the facility is listed in Tables 10 and 11 as a copper smelter.

It should be noted that these attributions ignore background (natural, regional, local -defined in footnote ⁷ on page 37) contributions to measured concentrations. This omission may be significant at some facilities that include major SO₂ emission sources that are not the subject of these assessments (e.g., the lead plant at Cominco-Trail). Details of the source attribution calculations are provided in CED (2000).

---

• 3. The electrowinning facility associated with the Boliden-Westmin Gibraltar Mines facility located at McLeese Lake, B.C., was not considered in detail. The operation was very small (annual copper production of about 2000 tonnes) and did not report any releases to air or water to the National Pollutant Release Inventory (NPRI). The plant ceased operation in 1999.
• 4. It should be noted that As is a metalloid rather than a metal. For simplicity, it will be referred to throughout this report as a metal. As will be noted in Section 2.4.1.1.3, however, there were circumstances under which As was necessarily handled differently than metals.
• 5. Some of the difficulties associated with assessing Hg using the critical load approach used in these assessments are discussed in de Vries and Bakker (1998).
• 6. While technically a VOC, CH₄ has been listed separately from other VOCs in Table 6. Unlike other VOCs, due to its negligible photochemical reactivity, CH₄ is not of significance in the formation of ground-level ozone or as a precursor in the secondary formation of PM. Methane is included here, however, as it is of significance to climate change.
• 7. Three "types" of background levels (ambient concentrations or deposition rates) must be considered in relation to atmospheric emissions in these assessments. "Natural background" refers to levels resulting only from natural sources. "Regional background" refers to levels that are considered typical of conditions over a large region such as the Canadian Shield. This may include some influence from distant anthropogenic sources as discussed in Section 2.3.1.2.2. "Local background" refers to levels in the immediate vicinity of the facilities that are due to all sources other than those being assessed. These values may be influenced by processes that are not the subject of these assessments, but which are being conducted at or near the copper or zinc processing facilities.